Cloud Computing: Start with the problem

At one point in my life I would build my own computing system for home use. I wanted a particular video card, a certain set of drives, and a lot of memory. Not only could I not find those things in a vendor’s pre-built computer, but those were more expensive – by a lot. As time moved on and the computing industry matured, I actually find that I can buy a vendor’s system as cheaply – and in some cases far more cheaply – than I can build it myself.  

This paradigm holds true for almost any product, even clothing and furniture. And it’s also held true for software…


If you need an office productivity package, you simply buy one or use open-sourced software for that. There’s really no need to write your own Word Processor – it’s kind of been done a thousand times over. Even if you need a full system for customer relationship management or other needs, you simply buy one.

But there is no “cloud solution in a box”.  Sure, if you’re after “Software as a Service” – type solutions, like being able to process video (Windows Azure Media Services) or running a Pig or Hive job in Hadoop (Hadoop on Windows Azure) you can simply use one of those, or if you just want to deploy a Virtual Machine (Windows Azure Virtual Machines) you can get that, but if you’re looking for a solution to a problem your organization has, you may need to mix Software, Infrastructure, and perhaps even Platforms (such as Windows Azure Computing) to solve the issue.

It’s all about starting from the problem-end first. We’ve become so accustomed to looking for a box of software that will solve the problem, that we often start with the solution and try to fit it to the problem, rather than the other way around. 

When I talk with my fellow architects at other companies, one of the hardest things to get them to do is to ignore the technology for a moment and describe what the issues are. It’s interesting to monitor the conversation and watch how many times we deviate from the problem into the solution.

So, in your work today, try a little experiment: watch how many times you go after a problem by starting with the solution. Tomorrow, make a conscious effort to reverse that. You might be surprised at the results.